Efficient on-demand spectrum sensing in sensor-aided cognitive radio networks

Abstract

Spectrum Sensing as a Service (SSaS) has recently emerged as a promising solution to realize dynamic spectrum access networks (aka, cognitive radio networks (CRNs)). This paradigm relies on a dedicated infrastructure wireless sensor network, operated by a spectrum sensing service provider S3P entity, for spectrum monitoring. Wireless users which are interested in knowing the status of a given spectrum band make a request to the S3P to perform that monitoring on its behalf. In this work, we study the problem of optimizing the selection of sensors needed to serve a given request such that the total cost (consumed energy) is minimized. We consider a spatial and temporal scheduling problem, in which, the energy levels at sensors, the maximum transmission range of secondary user (SUs), and the required monitoring time are taken into account. The problem is formulated as an Integer Linear Program (ILP). Then a sub-optimal greedy algorithm is proposed with two variations. Thorough evaluation shows that the proposed algorithm performs very well with respect to the optimal solution. Finally, results illustrate that the first algorithm variation is better in terms of energy cost, and the second variation allows for serving requests with a lower number of nodes.